1
Federalism and Constitutional Entrenchment
Jacob T. Levy1
The turn from the courts
One of the most striking developments in the past ten-plus years of
constitutional theory has been the partial or wholesale critique of judicial review
among those traditionally identified as ‘legal liberals’ or ‘liberal legalists.’ In its
moderate versions, this critique encompasses Cass Sunstein’s account of judicial
minimalism and Mark Tushnet’s call to “take the Constitution away from the
courts.”2 Its least moderate prominent version is the sweeping critique Jeremy
Waldron has offered over almost twenty years of most constitutional judicial
review in well-functioning democratic systems.3
These debates in recent political, legal, and constitutional theory about the
idea, and legitimacy, of constitutional entrenchment have mainly focused on the
entrenchment of substantive normative commitments and, especially, of bills of
judicially-enforceable rights.
American constitutional theory has been
Tomlinson Professor of Political Theory, McGill University, jacob.levy@mcgill.ca . Preliminary
draft; comments very welcome.
2 Cass Sunstein, One Case At A Time, Cambridge: Harvard University Press, 1999; Mark Tushnet,
Taking the Constitution Away from the Courts, Princeton: Princeton University Press, 2000. See also
Larry Kramer, The People Themselves, Oxford: Oxford University Press, 2005.
3 See, inter alia, Waldron, “A Right-Based Critique of Constitutional Rights,” Oxford Journal of
Legal Studies, Vol. 13, No. 1. (Spring, 1993), pp. 18-51; Law and Disagreement, Oxford University
Press, 1999; "The Core of the Case Against Judicial Review," 115 Yale Law Journal 1346 (2006). As
an aside, it seems to me that part of what makes Waldron’s challenge such a powerful one is that
he has taken what is perhaps the most important jurisprudential theory of this generation, Joseph
Raz’s analysis of legal authority, and done what Raz has never convincingly managed: joined the
jurisprudential theory to a normative political theory.
1
2
preoccupied with cases like these since at least the Warren Court; Canadian
constitutional theory has followed a similar path since the enactment of the
Charter of Rights and Freedoms; and the relevant debates in Britain have
centered on whether that country ought to have such a bill of rights.
But most entrenched constitutional provisions—which is to say, most
constitutional provisions in most states with written constitutions—concern
procedures and institutional arrangements. In this essay I try to draw attention
to such institutional arrangements, and particularly to the arrangements
surrounding federalism. While the constitutional entrenchment of federalism
and federalist arrangements is, I think, defensible and perhaps even
indispensable,4 it is not somehow outside the scope of the criticisms of
entrenchment that have been made in other contexts. Indeed, federalism is a
centrally important instance of the phenomenon under debate, and I do not think
we can understand the phenomenon without attention to it.
If, as I think,
I have developed my positive account of federalism and entrenchment in: “Federalism, Liberalism, and
the Separation of Loyalties,” 101(3) American Political Science Review, August 459-77, 2007; “Federalism and
the Old and New Liberalisms,” 24(1) Social Philosophy and Policy 306-26, 2007; "National Minorities Without
Nationalism," in Alain Dieckhoff, ed., The Politics of Belonging: Nationalism, Liberalism, and Pluralism,
Rowman & Littlefield 2004; "Indigenous self-government," in Nomos XLV: Secession and Self-Determination,
Stephen Macedo and Allen Buchanan, eds., New York: New York University Press, 2003; and "Literacy,
Language Rights, and the Modern State," in Will Kymlicka and Alan Patten, eds., Language Rights and
Political Theory, Oxford University Press 2003. I examine some consequences of decentralization without
entrenchment in “Three Perversities of Indian Law,” .forthcoming, Texas Review of Law and Politics, and
“Self-determination, non-domination, and federalism,” forthcoming, Hypatia (both available on SSRN).
The present paper is one of two planned to defend the theory developed across these papers against two
lines of criticism—the critique of constitutional entrenchment in this case, and the charge that the theory
requires violations of the publicity principle in the other.
4
3
entrenchment is legitimate with respect to rules of federation, then that may tell
us something about the legitimacy of the entrenchment of rights.
Throughout I mean to treat the critique of entrenchment as a serious and
powerful set of arguments.
For some years now I’ve suffered a bit of bad
conscience, knowing that I had no developed answer to that critique—especially
as it has been offered by Waldron—but being unwilling to embrace it and all its
implications. While this paper flows from my interests in federalism, and not
from a desire to overcome that bad conscience, I hope that it will be a first step in
answering that critique. Accordingly, much of the paper is devoted to showing
that federalist entrenchment does come within the critique’s terms, even though I
ultimately mean to defend rather than to damn such entrenchment. The paper
will also be especially concerned with Waldron’s critique of entrenchment, as I
think it ties together most of the themes in the “turn from the courts” literature
more generally.
The concepts in play
“Constitutional entrenchment” admits of variation; even within a given
constitutional order, some provisions may be amendable by weaker action than
others. I’ll refer to any rule or provision as entrenched which formally cannot be
altered by the same rules as ordinary legislation of the central government (i.e.
parliamentary majorities, with presidential acquiescence in presidential
4
systems).5 Such entrenchment can take the form of requiring amendment by
legislative supermajorities, by popular (majoritarian or supermajoritarian)
referenda, by the concurrence of center and provinces or of legislature and
electorate, or any combination of these.
But it is important to note that some
entrenched provisions are avowedly unamendable without discarding the
constitutional order altogether, such as the German Bill of Rights; and some are
effectively so, such as the rule guaranteeing equal representation for each state in
the United States Senate. Here I am not concerned with the differences among
levels of entrenchment, from slight-supermajoritarianism all the way to
supposed immutability.6
For purposes of this paper I will usually (but not always) elide one other
potentially important distinction concerning entrenchment: between the
entrenchment of a constitutional provision and its enforceability by means of judicial
review. That is, I will assume that entrenchment and judicial enforcement do go
Without trying to formalize the distinction, I’ll note that the assumption is that the entrenched rule is more
difficult to meet than that governing ordinary legislation—that is, what Adrian Vermeule refers to as
“submajority rules” are not included in the category. Vermeule, Mechanisms of Democracy.
6 Melissa Schwartzberg’s recent rich study Democracy and Legal Change, Cambridge: Cambridge
University Press, 2007, is centrally concerned with such cases of unamendability or
“immutability”. Accordingly, her only extensive discussion of federalist entrenchment comes in
the context of the guarantee that states’ equality of representation in the Senate cannot be altered
without their consent. She means for her critique of immutability to have implications for more
general cases of constitutional entrenchment—since, for example, she criticizes immutability for
moving the locus of legal change from the democratic legislature into the judiciary and she
praises the educational and deliberative benefits of a citizenry’s awareness that it can remake the
laws, both of which are normative positions that apply to questions of constitutional
entrenchment in general. But not all of her critiques spill over past immutability in that way, and
she leaves open the possibility that, e.g., some forms of supermajoritarian requirements for
constitutional amendment are compatible with her argument.
5
5
together—whether the enforcement is by a regular court or a specialized
constitutional one, whether it is retrospective or (as in France) prospective, and
whether the judiciary has the last word (prior to constitutional amendment) or
other actors may override the judiciary (as in the Canadian notwithstanding
clause).
A constitution that is nominally entrenched but declaratory in its
effects—one that tries to provide a focal point for public deliberation about
matters of foundational importance but that does not offer any institutional
restraints on the central legislature—falls outside the scope of this paper, even if
that declaratory constitution is entrenched in the sense of being difficult or
possible to amend. But, as we shall see, there are entrenched constitutional
provisions that are not merely declaratory but are also not, in anything like the
normal case, enforced by the judiciary.
The judicial override of legislative enactments does not exhaust the
actions that a constitution may authorize in its own defense, of course. I take it
that the core of constitutional entrenchment, indeed the core of the link between
constitutional rights and the judiciary, is the defense of procedural, rule-of-law
rights against executive action. Habeas corpus is fundamental to the rule of law
and thence to constitutionalism. Judicial limits on legislatures represent a kind
of late extension by analogy of the core judicial-constitutional function of
ensuring that persons are only imprisoned and punished in accordance with
6
enacted, prospective, promulgated laws.7
But this kind of thing is
uncontroversial among the theorists with whom I am concerned here; Waldron,
whose skepticism of judicial review is probably the most radical of the leading
participants in the debate, explicitly distinguishes between restraints on lawless
executive action and restraints on legislative lawmaking.8 By “constitutional
entrenchment” I refer only to the latter, taking the former for granted.
The most obvious subject matter for federalist entrenchment in particular is
the division of powers, authority, and responsibilities between the center and the
provinces—which of them has jurisdiction over what kind and scope of
questions, when and whether authority is shared concurrently between the
provinces and the center, and so on. In the United States, this is famously the
stuff of commerce clause jurisprudence, and the slight rejuvenation of federalist
constitutional enforcement under the Rehnquist Court centered on it.
But more fundamental is the entrenchment of the very existence of the
provinces—the question of whether the state is to be a federal one at all—and the
entrenchment of the identity, continuity, and borders of the provinces. The
formal allocation of authority is likely to be a dead letter if the center can
See, of course, Lon Fuller, The Morality of Law. See also: George Kateb, “Remarks on the
Procedures of Constitutional Democracy,” in The Inner Ocean, Ithaca: Cornell University Press,
1994; Judith Shklar, “Political Theory and the Rule of Law,” in Shklar, Political Thought and
Political Thinkers, Chicago: University of Chicago Press; Montesquieu, The Spirit of the Laws.
8 Waldron, “Torture and Positive Law,” Columbia Law Review
7
7
threaten to dissolve a recalcitrant province, or to gerrymander it into a morepliable shape, or to carve it up or lump it in with a larger neighbor. Closely
related is the guarantee of the provinces’ self-governing autonomy from the
center—the guarantee that a province’s government will be chosen by provincial
elections and not by central appointment.
(Again, the formal allocation of
authority can’t matter very much if the province is governed by an apparatchik
central appointee.) I’ll refer to the division of powers and the guarantee that
provinces have some independent legislative authority under the rubric of
provincial autonomy, and the guarantee of provinces’ existence, their immunity to
gerrymandering, and so on as pertaining to their integrity.9
Integrity is routinely guaranteed in the constitutions of federations. The
usual rule is that no province may be dismembered, diminished, or combined
with another without its consent—sometimes provincial-popular consent,
sometimes provincial-legislative. Argentina, Austria, Canada, Brazil, and the
United States all have such a constitutional provision. Germany’s provision is
complex but its basic outline is that changes to provincial boundaries require the
consent by referendum of both the populations of the whole states and the
populations of the affected territories.
Even quasi-federal Spain holds that
“Autonomy” and “integrity” are both terribly loaded concepts when discussing natural persons
and their freedom or identity. I don’t mean to import any of the baggage from those discussions;
here the terms are merely shorthand.
9
8
changes to provincial boundaries cannot be made by ordinary legislation,
requiring instead an “organic law” that must be passed with an absolute majority
of the lower house. India is the outlier in guaranteeing states no more than a
“voice;” Parliament may alter state boundaries with ordinary legislation. As far
as I can tell no other constitution usually thought of as federal allows such a
thing. (India is also an outlier in the weakness of its protection of autonomy,
with the possibility of state governments being dissolved by decree of the central
government.)
Conceptually quite distinct from either the allocation of regulatory
authority or the existence and integrity of the provinces are questions of the
provinces’ institutional participation in the government of the center—the status
of the German Bundesrat as the direct representative of the governments of the
Lander, the province-based representation in the U.S., Australian, Argentine,
Mexican, or Spanish Senate, or the U.S. Electoral College, and so on.10 I also
include in this category rules about general constitutional amendment that
depend on the provinces—three-fifths of state legislatures, a majority of the
voters in a majority of the states, and the like.
Sometimes constitutional
While the U.S. Senate obviously represents by states, it is sometimes overlooked that the U.S.
House does as well. The House does not represent population, but states on the basis of
population, with each state, no matter how small, guaranteed one representative, and only whole
numbers available. An American population of c. 298,000,000 divided by 435 representatives
yields about one representative per 685,000 constituents. But Wyoming has one representative
for its roughly 500,000 people, whereas Montana has one for more than 900,000.
10
9
amendments must first be approved by a bicameral legislature where one
chamber represents the provinces in some form, and then are subject to approval
by the provinces themselves or by the electorate divided provincially, so the
provinces participate in amendment both separately and jointly at the center.
(American amendments must be passed by the state-based Senate but then must
also be ratified by the states severally.)
Such provincial participation in the center and in constitutional
amendment is ubiquitous; I’m not sure that any federation lacks it entirely.
It
might be better understood as a mechanism for the protection of federalism
rather than as constitutive of federalism as such.11 If the provinces have no
independent legislative authority and no guarantees of their own stable
existence, then they are effectively local governments in a unitary state; direct
provincial participation in the government of the center is not similarly
constitutive of federalism.12 It might well be necessary to the effective stability of
a federal constitutional order—if the provinces lack any share of a governing say
See Herbert Wechsler, “The Political Safeguards of Federalism: The Role of the States in the
Composition and Selection of the National Government.” Columbia Law Review 54:543—560
(1954).
11
Here , as in many places in my work on federalism, I find that I have independently reached a
conclusion that Jenna Bednar had reached sooner and developed more in greater empirical
depth. Bednar treats “geopolitical division, independence, and direct governance [the ability of
both the center and the provinces to legislate over their constituents directly],” and the
entrenchment of these, as constitutive of federalism, but not “inter-level appointment.”
See
Chapter 1 of her forthcoming book, The Robust Federation, Cambridge University Press, 2008. On
the general topic of this paper, see Bednar, “Federalism, Judicial Independence, and the Power of
Precedent,” http://www-personal.umich.edu/~jbednar/WIP/precedent.pdf .
12
10
at the center, they may not be de facto able to protect their de jure integrity and
autonomy. Even if provincial participation at the center is not analytically
essential to federalism as such, I will treat its constitutional entrenchment as part
of the category “federalist entrenchment.” Provincial integrity, autonomy, and
participation are all characteristically constitutionally guaranteed in federations,
which in turn make up a large share of constitutional democracies in the world.
Federalist entrenchment is a conspicuous feature of entrenchment in general
Taken together, these forms of entrenchment distinguish federalism as an
intellectual category and constitutional phenomenon from other forms of
decentralization—that is, federalism in an important sense is entrenchment, and
if entrenchment is vulnerable to criticism then federalism likely is as well. 13
The critique(s) of entrenchment
Constitutional entrenchment in its current form dates to developments in
constitutional theory and practice in the American states in the 1770s and 80s.
Fundamental charters and written statements of founding laws were much older,
as was a so-called “ancient” constitutionalism which viewed the traditional and
inherited fundamental laws of a polity as binding on governors and beyond their
ability to alter. But the founding charters of the newly-independent American
11
states were recent creations, having none of the veneer of antiquity; and they
were typically created by the very legislatures they were meant to authorize and
legitimate. As Jefferson complained in the Notes on the State of Virginia, they tried
unsuccessfully to escape the reality and rule that no legislature could bind its
successor.
Even the procedure of supermajority rule for certain kinds of
lawmaking was, if passed by one legislature, vulnerable to repeal by a simple
majority of its successor. The innovations of constitutional conventions, outside
the legislature, to draft constitutions, and popular rather than legislative
ratification, seemed to solve these difficulties, legitimating supermajority rules
and binding constraints on legislatures.
When joined to the old but partly-
inchoate institutional idea that courts might enforce more fundamental laws
against legislative violations—that is, to institution of judicial review that
crystallized in the states in these decades—constitutional entrenchment took on a
roughly recognizable form. Some laws passed at time 1 by a special procedure
would be designated fundamental and superior to ordinary legislation and
executive action (and, so, constraining of such legislation and action).
This
superiority would authorize judges at time 2 to enforce the fundamental laws, as
laws, against other, illegal, government action. And these fundamental laws
could not be altered save by some combination of legislative supermajorities and
direct popular approval.
12
The key critiques of and complaints against constitutional entrenchment
are nearly as old as the phenomenon itself.
It is understood as antidemocratic
and countermajoritarian; as representing a kind of intergenerational tyranny; and
as improperly empowering the judiciary at the expense of the legislature. Note
that the first and third critiques are not identical. Madison was an advocate of
constitutional entrenchment on countermajoritarian grounds, but opposed the
tendency of judicial review to make the judiciary the supreme interpreter of that
entrenched constitution. One of the critiques might be blunted with an elected
judiciary, the other not; one is blunted by the existence of a partially-nonmajoritarian branch of the legislature (e.g. the U.S. Senate), the other not.
The
two critiques are, of course, related; in the final accounting Madison never could
reason out a way for the Constitution to constrain the legislature that didn’t in
effect mean that the courts could do so.14
But they are different in principle.
The
incompatible
countermajoritarian
critique
is
with
entrenched
constitutionalism as such. Even if a constitution could somehow be made selfMadison had proposed in the Convention that Congress as such review state laws for their
constitutionality, and that Congress’ own legislation be reviewed by a Council of Revision that
drew on the executive branch and the federal bench—in effect combining the subsequent
functions of presidential veto and judicial review. But in the Constitution as actually enacted,
while in 1788 he famously expressed the thought that judicial supremacy over the other branches
could “never be proper,” he came to understand that the joint operation of the articles declaring
“the articles declaring that the federal Constitution & laws shall be the supreme law of the land,
and that the Judicial Power of the U. S. shall extend to all cases arising under them” was to allow
“the Judicial Authority of the U.S. [to] be the constitutional resort for determining the line
between the federal & State jurisdictions.” Notably, this was after decades of conflict with a
Federalist-dominated judiciary that the thought had abused the trust; he insisted that “the abuse
of a trust does not disprove its existence.” Letter to Thomas Jefferson, June 27, 1823.
14
13
enforcing, it would constrain the choices made by democratic majorities. The
same is not true for the specific institutional critique of the judiciary.
Jefferson himself, insisting that “the dead have no rights,” held that it was
unjust for one generation to legislate over later ones. If few have taken up his
proposal that all laws (fundamental and otherwise) by repealed and institutions
restarted every generation, many more have thought that subsequent
generations ought to be free to reform, undo, or repeal inherited rules if they
wished.
That one generation could enact a rule that a majority of their
grandchildren could not repeal seemed to him, and has seemed to many since, a
kind of tyranny.
Moreover, since constitutional entrenchment (more-or-less) necessarily
empowers judges at the expense of legislatures, it has been vulnerable to attack
as antidemocratic, antimajoritarian, and so on.
Again, this is an enduring
complaint. Even Madison himself, though he embraced entrenchment and was
unable to solve the paradox of how the constitution could be legally superior to
legislation without giving too much power to the judiciary, was never
comfortable with that power. Insofar as the judiciary was made up of ordinary
human actors, and was not the Constitution incarnate, constitutional
14
entrenchment seemed to yield judicial review which seemed to make the
judiciary superior to the legislature, “which can never be proper.”15
As noted above, after some decades in which the main debates in
constitutional theory about judicial review and entrenchment seemed to concern
how they might best be defended and construed, judicial review has come under
serious academic skepticism in the past several years. Theorists and scholars
including Jeremy Waldron, Mark Tushnet, Cass Sunstein, and Adrian Vermeule
have
fleshed
out
and
refined
both
the
intergenerational critiques of entrenchment.
countermajoritarian
and
the
Waldron and Vermeule, for
example, both stress the idea of comparative institutional evaluation. They suggest
that the traditional case in favor of constitutional entrenchment with judicial
enforcement has come from a comparison of idealized judges with legislatures
that are viewed through much more gimlet eyes. A putatively unacceptable
legislative outcome is remembered or posited; the punch line is assumed.
“Surely you don’t think a legislature should be allowed”—passive voice—“to do
that?” But judges are just as real, just as human, actors as legislators—and just as
prone to errors. Philosophers and legal theorists may imagine themselves as
Herculean judges, sure that they would reach better, more-right answers on the
one or two questions of most concern to them than the real legislatures they see
15
in the world. But judges have limited time, limited information, limited moral
knowledge, and limited virtue, just like everyone else. Waldron stresses cases of
moral judgment and philosophical argument: rights, such as the right to freedom
of speech, have uncertain parameters and boundaries; they are the subject of
serious moral dispute among persons of good faith who believe in the rights.
Constitutional provisions about those rights are surrounded by interpretive
uncertainty, vagueness, or emptiness.
Given all of that, a five-out-of-nine
majority of the U.S. Supreme Court has no special claim to expertise or certainty
compared with a majority of democratically elected representatives offering a
competing interpretation of the right in question.
Vermeule tends to stress
questions of technical knowledge and expertise, in which the issues are less
epistemic and more technocratic, but the upshot is more or less the same: reallyexisting judges are poorly situated to substitute their judgment for that of
legislatures.
And, according to the critique of entrenchment, we must always
remember that it is judges, not the constitution, that will be limiting the legislature.
A constitution is not self-interpreting or self-enforcing, and there is no way to
make it so. This means that we never face the abstract question: should it be
unconstitutional for legislatures to do X? but only the mundane question: would
16
it be better for judges to be able to decide, in the name of preventing X, whether
and when to block legislative action?
To all of this Waldron in particular adds a substantive normative defense
of democratic legislation and legislative decisionmaking.
He intends his
comparative institutional evaluation to include a comparison between best-case
legislatures and judiciaries (whereas, e.g., Vermeule is primarily concerned to
compare really-existing legislatures and judiciaries; they share a sense that
Dworkin and his intellectual kin have mainly compared best-case judiciaries
with really-existing legislatures). In particular, he argues that there is something
deeply normatively appropriate about democratic majorities acting through their
legislatures to decide contested questions of rights and rights-interpretation. He
stresses that this is not, as it has traditionally been viewed, simply a capitulation
to majority tyranny and majoritarian deprivation of minority rights:
The point to remember here is that nothing tyrannical happens to me merely by
virtue of the fact that my opinion is not acted upon by a community of which I
am a member. Provided that the opinion that is acted upon takes my interests
properly into account along with everyone else’s, the fact that my opinion did
not prevail is not itself a threat to my rights, or to my freedom, or to my wellbeing. None of this changes necessarily if I am also a member of the topical
minority whose rights are at issue.
Instead, he suggests that the same image of individuals as autonomous moral
agents that underpins liberal rights theory also requires respecting our fellow
citizens as interpreters of rights—and so respecting the legitimate claims of
17
majorities of persons to settle arguments about such interpretations. We will
return to these points below.
Federalism and the critiques
Federalist entrenchment in all its forms runs afoul of at least some of the
critiques.
Ruth Gavison, in one of the only articles to broaden the question of
constitutional entrenchment beyond questions of individual rights, holds that
“Only unitary states have the luxury (or burden) of debating the question
whether governmental structure should be included in the constitution. All states
must have provisions regulating these subjects, but in unitary states these may
evolve, and be enforced in part as constitutional conventions. This is precisely
what has happened in England, and Dicey is still a powerful spokesperson for
the desirability of this way of regulating the structure of government. However,
federal governments cannot make this choice, since it is a central issue of such
governments that the division of powers between sub-units and the central
government is determined in a way that will be beyond the unilateral change of
either states or the central government.[…] If the state is a federal one – the
relations between central and local government must be made in a formal
constitution, which cannot be changed unilaterally by the member parts.”16
Now Gavison is not obviously right about the relationship between federalism
and written constitutions.
If relations among various branches of a single
government can be regulated by evolving, unwritten, constitutional customs and
conventions with occasional acts of legislation—as has been the case for relations
Ruth Gavison, “What Belongs in a Constitution?” 13 Constitutional Political Economy 89-105
2004.
16
18
among Commons, Lords, cabinet, and Crown in Britain—then relations between
a center and provinces might well be regulated the same way. Indeed, Great
Britain itself has never been quite the simple undifferentiated unitary state
imagined by Dicey even before the recent decentralizing reforms.17 Unwritten or
unentrenched constitutional norms governed Anglo-Scottish relations before the
Act of Union; such norms plus the Act itself governed the continuing
constitutional differences between England and Scotland (in their legal systems,
established churches, banking and university laws, and so on) afterward; and the
relationship of Ireland to Westminster was always complicated but hardly ever
simply that of being part of the territory of a unitary state governed by a unitary
legislature. (And this is to say nothing of even-more-complex cases such as the
Channel Islands.)
Constitutional divisions of power between center and
provinces predated written constitutions throughout Europe; and if the so-called
“ancient” constitutions of France and Spain more or less collapsed into de facto
unitary state even before the Revolution, the Holy Roman Empire endured as a
kind of federation without a modern written constitution for centuries, and the
Dutch and Swiss confederations had written treaties at their foundings but were
partly governed by unwritten constitutions thereafter.
Note Gavison’s telling mistake of referring, in this context, to “England” rather than “Great
Britain” or “the United Kingdom.”
17
19
In the modern world, so few states lack written constitutions that it is
difficult to get any traction on the question of what can and what cannot happen
under an unwritten-constitutional order. Every modern democratic federation
has a written constitution; but so does every modern democratic state besides the
U.K. and Israel.
Every modern democratic federation has an entrenched
constitution, but so does every modern democratic state besides the U.K., Israel,
and New Zealand. Given Britain’s unique institutional continuity and Israel’s
and New Zealand’s tiny sizes, I doubt that we can safely generalize about the
conditions that would allow a modern liberal democracy to persist without a
written or entrenched constitution.
The question of federalist entrenchment
cannot be short-circuited so easily.
Gavison moreover tries to treat institutional entrenchment generally as
outside the scope of the debate.
“It is important to note that, in distinction with the situation concerning rights
and credos, there is no serious argument against judicial review of the
‘institutional’ part of constitutions. This consensus is based on the fact that the
provisions of the constitution in these matters are relatively clear, and that there
is a necessity that there will be an authoritative arbiter of the disputes that do
arise.” (93)
But this won’t do either. There is and has been tremendously serious argument
against judicial policing of the institutional boundaries between the legislature
and the executive in the United States, and the judiciary itself has often been
extremely reluctant to act as that authoritative arbiter. (The “provisions in the
20
constitution in these matters” are anything but “relatively clear”!)
More
appositely, the question of whether the American federal judiciary could police
the allocation of authority between the center and the states has been a terribly
vexed one in American constitutional history; until the Warren Court’s rights
revolution, it was the core of debates about judicial review. Since 1937 the
dominant view has been that it is better for Congress to unilaterally decide what
lies within the scope of its power to regulate interstate commerce and what lies
outside of it than for the courts to do so. That this doctrine’s edges were nibbled
a bit by the Rehnquist Court does not change its real force. In Raich v. Ashcroft,
the question at stake was not really whether the non-commercial growth of
marijuana for intrastate personal use qualified as “interstate commerce,” but
rather whether it lay within Congress’ authority to decide whether it so qualified. The
Supreme Court opted for deference to Congressional authority, which is to say
that it declined to genuinely review the question of whether the boundary
around “interstate commerce” had been crossed or not.
The “authoritative
arbiter” remains Congress, judging in its own case about how far its power
extends and how little legislative autonomy is left to the states.18
I describe it this way advisedly, remaining mindful of Waldron’s powerful Hobbesian-cumRazian complaint that “Almost any conceivable decision-rule will eventually involve someone
deciding in his own case. Unless we envisage a literally endless chain of appeals, there will
always be some person or institution whose decision is final. And of that person or institution,
we can always say that because it has the last word, its members are ipso facto ruling on the
acceptability of their own view. Facile invocations of nemo iudex in sua causa are no excuse for
18
21
It seems to me that we are left with no choice but to agree with Waldron’s
suggestion on this point:
“[M]any of the challenges to rights-oriented judicial review can be posed to other
forms of constitutional review as well. In recent years, for example, the Supreme
Court of the United States has struck down a number of statutes because they
conflict with the Supreme Court’s vision of federalism. Now, everyone concedes
that the country is governed on a quite different basis so far as the relation
between state and central government is concerned than it was at the end of the
eighteenth century, when most of the constitutional text was ratified, or in the
middle of the nineteenth century, when the text on federal structure was last
modified to any substantial extent. But opinions differ as to what the new basis
of state/federal relations should be. The text of the Constitution does not settle
that matter. So it is settled instead by voting among Justices—some voting for
one conception of federalism (which they then read into the Constitution), the
others for another, and whichever side has the most votes on the Court prevails.
It is not clear that this is an appropriate basis for the settlement of structural
terms of association among a free and democratic people.” (2006: 1357)
Moreover, the entrenchment of the autonomy and the integrity of provinces both
carry important costs above and beyond those associated with the entrenchment
of rights.
That is, federalist entrenchment may stand in greater need of
justificatory work than the entrenchment of rights, not lesser. The wide variety
of public goods, regulations, and policies would seemingly be most efficiently
provided by similarly-various levels of government. And social, economic, and
technological change over time probably changes the level at which some
policies are most-efficiently provided. Provincial boundaries often fail to match
forgetting the elementary logic of legitimacy: People disagree, and there is need for a final
decision and a final decision procedure.” (2006: 1400-01). The only way to prevent Congress
judging in its own case how far its powers extend is to leave the Supreme Court endlessly
judging in its own case about the scope of its own authority.
22
up with either economic regions such as metropolitan areas or environmental
ones such as watersheds; in both cases the best policy-fitting region may include
parts of several provinces but not all of any province. Federalist entrenchment
picks out provincial units that may not be the right size for any particularly
important kind of policy, and grants them constitutional status which is denied
to, e.g., cities, counties, metropolitan regions, and watersheds. Entrenchment
moreover freezes a particular allocation of authority between provinces and the
center that may well become inappropriate as time goes on. The alternative to
federalist entrenchment, in short, need not be unitary centralized government; it
might be decentralized government with more flexibility about what levels of
government performs which functions when.
The difficulties involved might greatly exceed fiscal or regulatory
inefficiency. As the cliché has it, the oak that does not bend may break, whereas
the willow’s lack of rigidity can prove to be a strength. Consider the attempts to
stave off constitutional crisis in Canada by the Charlottetown and Meech Lake
Accords—what Michael Lusztig has called the “constitutional paralysis” of the
Canadian federation.19 The status quo is sufficiently entrenched that it seems
impossible to reform the constitutional order in ways that would make it more
See Michael Lusztig, “Constitutional Paralysis: Why Canadian Constitutional Initiatives Are
Doomed to Fail,” Canadian Journal of Political Science / Revue canadienne de science politique, Vol. 27,
No. 4. (Dec., 1994), pp. 747-771.
19
23
palatable to Quebec. The final 1992 breakdown of reform efforts that would
recognize Quebec as a “distinct society” and formalize asymmetric federalism
came within a hair’s breadth of prompting Quebecois secession in 1995.
If
entrenchment offers certain kinds of guarantees and stability, it by the same
token yields inflexibility. We have seen that some federalist entrenchment, such
as the allocation of authority between provinces and the center, is prone to the
same interpretive difficulties as the entrenchment of rights. But those federalist
rules that aren’t prone to interpretive uncertainty may entrench procedures,
institutions, and distributions of power and authority that eventually become
unacceptable to some province(s) that have the capacity to break the state but
that are not effectively reformable—a set of risks not ordinarily associated with
the entrenchment of rights.
On the other hand, there is a sense in which federalist entrenchment
seems less vulnerable to democratic or majoritarian objections than does the
entrenchment of substantive rights. The entrenchment of provincial integrity
and autonomy, at least, act as guarantors of provincial democratic or majoritarian
authority. They may empower the judiciary to act against a national legislature
when it, e.g., wishes to dismember a province or legislate on a matter reserved to
the provinces; but they do so in order to leave the provincial demos/ government/
legislature with the final say. In the case of judicial enforcement of substantive
24
rights, it appears to be the judiciary that retains the final say on what the
government may do; federalist entrenchment requires only that the judiciary
referee the boundary between provincial and central elected governments. And
so, for example, Melissa Schwartzberg’s defense of the educative virtues of
democratic decisionmaking about fundamental questions does not decisively tell
against federalist entrenchment.
The entrenchment of provincial participation at the center does run afoul
of democratic objections; indeed, perhaps the most non-majoritarian feature of
national governments in many federations is the equal representation of each
province in one chamber of the central legislature.
Moreover, many have
thought that the defenses of autonomy and integrity were also undemocratic. In
this, though, they were mistaken; they relied on a typically unargued-for
Rousseauian premise that the national-level demos was the true one that had an
authentic democratic will, while the peoples of the several provinces were
somehow less democratically real. The idea that a national-level people has an
authentic pre-institutional will.
The difficulties with assuming an authentic
democratic will that precedes and exists outside of institutional forms and
procedures will come up again in the next section.
Constitutions, procedures, and institutions
25
Constitutional entrenchment has often been defended as a kind of
precommitment strategy: the people, like Ulysses, bind themselves (or rather
“itself;” conceiving of the people as a corporate actor matters for the coherence of
the view) against a predictable future temptation, passion, or weakness of will.
Precommitment views are meant to be relatively immune to the charge of
countermajoritarianism or anti-democracy, since the constitutional constraint is
both self-imposed (the constitution being democratically enacted) and
autonomy-enhancing rather than autonomy-limiting (the principle being the
more fundamental popular will than the exception). Waldron, unsurprisingly,
dismisses all versions of precommitment defenses of entrenchment. 20 Some of
his critiques are well-taken; the people don’t make up a unified actor at the time
of the enactment of the principle; later disagreements about interpreting the
principle may well be good-faith arguments, not evidence of panic or weakness
of will; a majority vote of a judicial body is not obviously more capable than a
majority vote of a legislative body of correctly divining the meaning behind the
original restraint; and a constitutional restraint can’t meaningfully be understood
as “self-“imposed à la Ulysses and the Sirens when it binds later generations.
Following an account of disagreement at the time of constitutional enactment, disagreement
about constitutional interpretation at the time of legislation, and that disagreement reappearing
among judges at the time of adjudication, Waldron writes, “my theme in all this is reasonable
disagreement, but I cannot restrain myself from saying that anyone who thinks a narrative like
this is appropriately modeled by the story of Ulysses and the sirens is an idiot.” LD, 268.
20
26
Waldron distinguishes between precommitment mechanisms that rely on
external causal mechanisms—say, Ulysses handcuffs himself and throws the key
overboard—and those that rely on the judgment of another actor, such as
Ulysses’ crew. I might purchase a mechanism that disables my car from starting
if my breath reveals excessive alcohol consumption; or I might give my keys to a
trusted friend. Waldron notes the obvious disadvantage of mechanical models:
they do not admit of exceptions. (It might be better to drive while slightly drunk
than to refrain from driving a sick child to the hospital.)
He continues:
“Clearly, if constitutional constraints are regarded as forms of democratic
precommitment, they operate more on the model of the friend’s judgment than
on the model of a causal mechanism. Except in rare cases (like ‘dual key’
controls of nuclear weapons) constitutional constraints do not operate
mechanically, but work instead by vesting a power of decision in some person or
body of persons (a court), whose job it is to determine as a matter of judgment
whether conduct that is contemplated (say, by the legislature) at t2 violates a
constraint adopted at t1.” (LD 261-2)
This is revealingly wrong.
The disadvantages of mechanistic precommitment strategies are matched
by an advantage: they are very reliable.
If they guarantee that there will
sometimes be Type 2 errors (false negatives, times when an action is disallowed
that should be allowed), they also guarantee that there will be no Type 1 errors
(false positives, when an action is allowed that should be disallowed). Which is
preferable, and whether either is preferable to refraining from precommitment, is
27
entirely dependent on the probability, frequency, and consequences of these
errors, not on any a priori rule. For any heavy drinker who does not, e.g., have a
child in the house who has a condition that predictably and regularly needs lifesaving emergency room treatment, I suspect that the breathalyzer in the steering
wheel is a much better precommitment strategy than giving keys to friends who
might themselves get drunk, or who might be susceptible to persuasion,
emotional blackmail (Ulysses’ curses and threats), or even drunken violence to
give keys back. Trusted friends may make mistakes in judgment; they may also
be prevailed upon to relax their vigilance, one way or another. So far this seems
to tell in favor of Waldron’s broader argument; if the demos really were like a
passionate drunk, willing and able to run roughshod over minorities, then it
might also be willing and able to run roughshod over the judiciary or the
constitution.21 For judgment-based precommitment to be a sound strategy, a
principal (P) at t1 must have reason to think both that the agent’s (A) judgment at
t2 will be much better, by P’s own t1-lights, than P’s; and that the reasons for P’s
impaired judgment at t2 aren’t so likely to lead P to induce A to give over as to
neutralize the advantage in judgment.
And, of course, in practice it has been very difficult for judiciaries to resist democratic
majorities in just those cases that we’re most prone to see as analogous to drunken panics: the
Red Scare, Korematsu, etc.
21
28
But if mechanical precommitments can be found, then the quality of
decisions might rise. True, there will be few cases in law or politics that are
strictly mechanical, on analogy with handcuffs and keys thrown into the ocean.
But there are plenty of such cases in which the constraint operating at t2 is not
merely some agent with temporarily-better judgment on the merits of the
question. The U.S. Constitution prohibits expenditures for a standing army more
than two years in advance—this is a precommitment.22
It denies to future
Congresses the ability to do something that they might deem preferable on the
merits when the time comes.
And while one could imagine a case being
litigated, that litigation would not just be a request for judges to decide whether
a longer-term appropriation was desirable on the merits.
Waldron’s focus on bill of rights-style cases has misled him here. Most
constitutional rules aim to improve decsionmaking relatively mechanically, not
through mere agent-substitution.
Agent-substitution asks someone else to
decide the merits of the contested question; relatively mechanical solutions aim
to alter the question, procedurally avoid it, or make some outcomes rather than
others more likely by giving actors with predictably different interests or
opinions some power over it. The rule against long-term military appropriations
At least it is as much of a precommitment as any of the other standard cases such as the First
Amendment, even though neither it nor the standard case involves an agent (the pouvoir
constituante) literally binding itself; but the gap between the legislature and the electorate can’t
be very significant for Waldron, since legislative majorities are morally legitimate ways to resolve
popular disagreements on his account.
22
29
prejudges particular disagreements, substituting relatively mechanical questions
(two years or less? Standing army?) for the substantive one. If other agents are
called in, it is to settle the mechanical questions, not the substantive ones; and for
the most part other agents aren’t called in at all.
This is all done because the outcomes of some kinds of errors are
considered worse than the outcomes of others, or because some errors are
deemed likelier than others, or some combination of these. In this case, rather
than encouraging legislators to decide the merits of the question each time, the
constitutional designers opted to fend off the risk of a legislature mistakenly
authorizing a permanent standing army under the control of a no-longeraccountable executive. That increases the risk of some other kinds of errors, of
course—perhaps not only the risk of under-funding, but also the risk of
misallocations of funding because some military allocations are made
inefficiently frequently.
Elsewhere Waldron critiques this kind of outcome-
oriented constitutional design as well.
“But a citizenry who disagree about what would count as the right results are not
in a position to construct their constitution on this basis… Using a result-driven
approach, different citizens will seek to design the constitution on a different
basis.” (LD 294)
This, it seems to me, proves too much—perhaps vastly too much. Its
intuitive power is blunted by all of our experience of institutional and
constitutional design. We cannot get very far in designing any decision-making
30
procedures, democratic or otherwise, with no reference to results. Arguments
about procedural fairness as such run out much too quickly, and anyway seem
insufficient. Constitutional designers and ratifiers have, and perceive themselves
to have, some level of Weberian responsibility for attention to consequences.
They want to create institutions that will enable a state to defend itself effectively
without being dominated by its own military, for instance—a pair of concerns
that runs right through The Federalist.
Waldron means to rule out, e.g., the libertarian move of assuming that the
common-law baseline of contract and property law is basically just and efficient,
and therefore proposing procedures that simply slow and impede legislation in a
general way in order to prevent economic regulation. This, he thinks, violates
respect for other persons as rights-interpreters, or as bearers of the Arendtian
“right to have rights.” We disagree about what rights persons as owners and
contractors and consumers and producers have; and that disagreement is the
stuff of politics among persons who have the right to argue and interpret
questions about rights. And surely there is something correct here. Procedures
meant to help us manage our disagreements should not beg the question and
take a side in those disagreements. When we have agreed to flip a coin to settle a
question, if I were to call heads and then say, “let’s use my weighted coin; it
comes up heads 60% of the time, and my view is right,” I would have missed the
31
point of the procedure. It can’t be a good reason to adopt an electoral rule that it
would help one’s preferred party.23
But there that can’t be all there is to say. Shared political life isn’t like
coin-flips all the way down. Even in the face of good-faith disagreement about
ends, institutions are designed with an eye toward the ends they will promote as
a matter of course. We try to avoid designing them to get particular outcomes in
particular cases; but we choose electoral rules to promote two-party stability or
multiparty
representativeness;
judicial
rules
to
promote
accuracy
or
overprotecting the innocent; faculty hiring rules to encourage rapid hiring or
waiting for ideal candidates or reinforcing expertise or building breadth. And
we do so knowing that we will not only disagree about particular cases but that
we also disagree about the balances between or among these various
considerations.
To put it a different way, in the face of disagreement about particular
cases, and disagreement about the reasons for deciding cases in general, we do not
simply adopt the decision rule to decide each case one at a time, or to re-argue the reasons
each time. We do some of each of those things; the procedures we adopt are not
carved in stone, and they come under pressure when they seem to generate a
It is, of course, an entirely common reason for doing so, and this turns out to be tolerable in a
relatively balanced party system; democracies don’t collapse as a result. But the rampant
hypocrisy around this is telling.
23
32
sufficient number of wrong outcomes. But we also devise rules to keep any
number of questions from being considered, and other rules that tend to favor
one or another of the contested reasons. We overweight the avoidance of some
kinds of mistakes that we (or a majority of us at the time of rules-adoption)
consider more serious, or harder to undo, or more likely to engender further
mistakes in the future.
In constitutional theory about entrenchment and judicial review, it may be
that no line between substance and procedure will be particularly stable, because
procedures are adopted in order to advance substantive outcomes; our
proceduralist commitments themselves will always be underdeterminate with
respect to the shape of institutions. The line could only be stable if one believed
in a preinstitutional Rousseauian general will, which Waldron (rightly) does not.
Otherwise the democratic will is always in part a product of procedures, not
something neutrally or transparently transmitted through them.
So we will
always have process-defining constitutional provisions (legislatures cannot
bootstrap their own legitimacy), and the choice among them will, at least very
often, be driven by substantive ends. While Waldron doesn’t need the same
substance-procedure distinction that, for example, John Hart Ely does—he does
not think that substantive and procedural rights fall into different categories—he
does need a fairly stable distinction in order to defend the purity of his
33
democratic-legislative procedures from smuggled-in substantive ends. Once it is
thought legitimate to frame procedures with an eye on their substantive
outcomes, overweighting against some kinds of errors and not others, then it
seems that the freedom of majorities to interpret contested rights-claims free
from interference by other agents is no trump.
Federalism as precommitment and substantive procedure
Federalist entrenchment offers the sort of institutional rules described in the last
section. Assuming that we may elide the distinction between the people as
constitutional enacters and the later legislative representatives of later people, it
is a kind of precommitment—not a precommitment to avoid madness or
drunkenness or running a ship aground on the rocks or any of the traditionally
colorful metaphors invoked to make precommitment an opposition between
reason and unreason, but a relatively mechanical set of decision rules, adopted to
affect the likelihood of some outcomes rather than others.
Federalist
entrenchment prevents later state-wide majorities from making certain kinds of
changes to the decentralization of the society’s governing institutions. It weights
34
the prevention of excess centralization more heavily than it weights the risk that
needed centralization will be resisted.24
Why not allow the central legislature to decide for itself, on an ongoing
basis, how much decentralization is needed, and where the boundaries should
lie?25 I think that part of the answer lies in path dependence. To a substantial
extent Waldron and Vermeule alike treat fundamental political decisions as
independent from each other; one decision one year, another the next. The same
is true, e.g., of Schwartzberg’s emphasis on the learning that can come from a
democracy’s correction of its own errors.
Errors are inevitable, but that is true no matter what we do, so pointing to
legislative errors is no reason to limit legislative authority. But in some areas of
political life it’s unlikely that serial decisions can be wholly or even mostly
independent of one another—the precommitment to avoid long-term military
This is a simplification and shorthand; many of the entrenched constitutional rules about
federalism limit provinces and empower the center. But these are more rarely the objects of
constitutional controversy except to determine where they run out (e.g. the U.S. Commerce
Clause). In the case of federations formed out of preexisting polities—the U.S., Canada,
Australia—the constitutional limits on provinces may be the more important constitutional rules
in early decades, as the provinces learn to live with their reduced status. But this tends to change
over time; eventually, it is the center rather than the provinces that is more likely to overreach.
25 Remember that in general federalist entrenchment does not predetermine policy disputes (what
regulations shall we have?) but rather rules about determining such disputes (what level of
government may decide what regulations we will have?) Again, this is not substantively neutral,
but it is a proceduralist precommitment about procedures, one step removed from questions of
substantive policy. In short, federalist entrenchment is more concerned with limiting the ability
of democratic majorities to decide who will decide various questions more than it is with directly
limiting their decisions about the questions; it is this that makes it a relatively mechanical
precommitment.
24
35
expenditures rests on the thought that an error in one direction (too much
funding of too strong an army under too unconstrained an executive) might have
long-term institutional consequences that make it uncorrectable.
If federalist entrenchment guards against errors that are less dramatic, it is
nonetheless similar in kind. It rests on the thought that the inevitability of error
in deciding the details of decentralization is not good enough reason to leave
such decisions up to new case-by-case decision every time. Some decisions
would have long-term consequences making themselves difficult to correct; and
some kinds of errors are more likely than others.
Some degree of
decentralization is desirable, but unconstrained central legislatures may well
decentralize too little. And if they do, the error will snowball, in the face of the
powerful tendencies in modern states toward centralization. To identify just a
few reasons for this:
1)
the center, as the usual site for armed service and wartime loyalty, tends
to attract a permanent degree of political attachment from those who have served
in the military and periodic increases in attachment during wars and
international crises, undermining the energy and enthusiasm citizens have for
maintaining the potentially-oppositional provinces;
36
2)
the center’s typically greater command of resources give it the ability to
coopt or corrupt provincial officials, or simply to undermine their policy
autonomy with selective expenditures;
3)
the center’s (ex hypothesi) unconstrained ability to alter the boundaries of
or abolish provinces both leaves the provinces as too unstable to engender citizen
loyalty and as vulnerable to manipulation by temporary national majorities,
jeopardizing the ability of the provinces to remain oppositional. In short, the
absence of federalist entrenchment risks making decentralized levels of
government partisan playthings on the model of Congressional districts in many
U.S. states.
Waldron has for years offered a critique of the “I expect you’d like to
know what I would do if I were a philosopher king or a Supreme Court justice”
mindset of political philosophy.26
The critique is a powerful one—more
powerful, I think, that he recognizes. There is a parallel critique to be leveled at
the idealization of uniformity and fixity in legislation; the legislature like the
philosopher may show an unhealthy impatience for letting things unfold over
time, and may suffer from a hubristic sense of having all the answers already in
hand.
Of course, judges may suffer from this error, too—this is what ties
Waldron’s methodological critique of much applied normative political
In addition to all the works on constitutionalism already cited, see his “What Plato Would
Allow,” in Nomos XXXVII: Theory and Practice.
26
37
philosophy to his critiques of Dworkin and of judicial review. But federalist
entrenchment only empowers judges at a couple degrees of remove.
In the first instance it tends to empower provincial legislatures, allowing
both for the operation of a jurisdictional competition that can operate as a
market-like discovery procedure for best policies, and for a peaceful coexistence
of disagreeing political groups without always requiring that disagreements be
settled by unified majority opinion. Philosophers may presume too much when
they think that they can settle deeply contested questions of rights and justice in
ways that should be binding on societies filled with Arendtian rights-bearers.
But the representatives of those rights-bearers may presume too much, too; they
may be too impatient of disagreement or of the time required for jurisdictional
competition to result in discoveries. The wisdom of the multitude 27 that Waldron
contrasts to the hubris of the judge is not found only in large numbers of voters
reaching one-off decisions. It may also be found in processes.28 And, when we
compare the institutional option of a uniform central all-powerful legislature
with the institutional option of legal competition, rather than with the
institutional option of an all-powerful judiciary, we see that virtues Waldron
“The Wisdom of the Multitude,” 23(4) Political Theory 563-84, 1995
See, inter alia, Hayek, `The use of knowledge in society'. American Economic Review , vol. 35, pp.
519-30 and Hayek, “Competition as a Discovery Procedure;” the link between Hayekian
discovery procedures and information-aggregating processes more generally has been explored
by Cass Sunstein, most prominently in Infotopia, Oxford: Oxford University Press, 2006.
27
28
38
claims for democratic processes may be found in other processes as well. And
federalist entrenchment may protect these processes.
It may be particularly desirable in the frequent case of ethnocultural
federalism.29 There is a deep consistency between Waldron’s democratic critique
of constitutionalism on one hand and his critique of multiculturalism, special
consideration for indigenous peoples, and treaties on the other.30 Indeed I think
they are two strands of an argument that Waldron had formulated almost twenty
years ago, with his “Rights and Majorities: Rousseau Revisited,” an argument
that majoritarian decision-making was fully compatible with respect for the moral
rights of minorities, and that what had traditionally been viewed as
entrenchments of minority rights should be re-understood as claims for excessive
minoritarian decision-making power over what all will do. In addition to a
defense of majoritarianism, the two strands share a present-and-future
orientation. Decisions about how we are to live are decisions we must make, now,
and the dead have no rights to control them, whether through the Constitution of
the United States or the Treaty of Waitangi or the original act of federation
between Quebec and Ontario. We should make them (at least mainly) with
I have attempted to show that ethnocultural federalism should be understood as part of the
normal case of federalism, not as a strange exceptional one, throughout the papers cited above.
30
See Waldron, “Minority Cultures and the Cosmopolitan Alternative;” “What is
‘Cosmopolitan’?”
“Superseding Historical Injustice,” “Cultural Identity and Civic
Responsibility,” “Indigeneity: First Peoples and Last Occupancy,” and “Redressing Historic
Injustice.”
29
39
respect to the rights and interests of those living now and those to come, not with
respect to either the agreements reached or the injustices committed by our
ancestors.31
But in cases of ethnocultural divisions between majorities and minorities,
errors on the part of majoritarian procedures may be both especially likely and
especially difficult to reverse. To think this we need not think that majorities are
intentionally rapacious or bigoted or vicious, or that minorities have as a matter
of moral or constitutional fact the rights that they think they have—the views of
“majority tyranny” that Waldron has disputed so vigorously. We need only
think that majorities are prone to ordinary kinds of mistakes for ordinary
reasons: they know their own circumstances best and generalize from them, their
opinions about hard moral questions are subtly influenced by their interests, and
so on.
The mistakes majorities can make in ethnoculturally divided are
moreover especially likely to spiral and snowball—a bit of unjust assimilation
makes living in the minority language that much less appealing to the younger
Here and in what follows, I am offering an implicit rejoinder to the view laid out by James
Tully, Strange Multiplicity, that the modern constitutionalism of the French and American
Revolutions and the two centuries since is deeply incompatible with respect for minority cultural
rights. That rejoinder is explicit in my “Montesquieu’s Constitutional Legacies,” in Rebecca
Kingston, ed., Modernity in Question: Montesquieu and His Legacy, SUNY Press, forthcoming, and
“Constitutions Without Social Contracts, Political Theory, forthcoming. Like Waldron, I think that
the affinities between countermajoritarian constitutionalism and multiculturalism are more
fundamental than the tensions between them—though I am sympathetic to both and he is
sympathetic to neither.
31
40
generation than it would have been, a bit of excessive centralization of
decisionmaking authority undermines the political institutions and authority
that the minority might have used to check the next bit of excessive
centralization, a bit of policymaking that privileges majority interests a bit too
much alienates and radicalizes the minority and hardens distrust between the
two groups. I think that all of this is related to the fact that almost every peaceful
and stable multinational democracy is federal to some substantial degree;
federalism is surely not sufficient for multinational democratic coexistence, but
may be necessary, because where it is absent the majority tends to centralize
decisionmaking in a way that is self-reinforcing and eventually makes things
more or less intolerable for the minority.
The category of “relatively mechanical” entrenchment is a
deliberately broad one.
In the case of federalist entrenchment, provincial
participation at the center is very mechanical. Provinces are given a kind of
corporate political power; electorates are organized in provincially-centered
ways. I suppose that the guarantee of state equality in the Senate is justiceable,
as is the existence of the Electoral College—if Congress purported to alter either
by ordinary legislation, I suppose that courts would intervene. But the scope for
interpretive uncertainty is so limited that it’s hard to imagine Congress trying to
do that—short of a moment of revolutionary constitutional politics in which the
41
federal courts’ say-so would probably not much matter. The entrenchment of
provincial autonomy is also very mechanical, admitting of very little interpretive
uncertainty and therefore inviting very little judicial oversight.
There is
uncertainty, again, in revolutionary moments—as in the revolutionary secession
of West Virginia from the revolutionarily-secessionist Virginia.
In general,
however, the rule is so mechanical as to shape the political framework and the
expectations of political actors, and not to become an object of interpretive
contention within politics. These very mechanical entrenchments are successful
precommitments. If they are normatively justified at all, then it is justified to
entrench them, not to prevent drunken or Ulysses-like majorities from overriding
them in fits of passion, but to prevent the basic rules of politics from constantly
being renegotiated or re-gamed.
The example of American Congressional
gerrymandering—once every decade, whichever party happens to be in control
of a state’s government shamelessly grabs for a maximally large share of that
state’s Congressional representation—as to remind us of the virtues of even an
arbitrary stability.
The entrenchment of provincial autonomy32 is only relatively mechanical.
What is entrenched is not a procedure or an institution, but a rule about what
32
I have almost entirely discussed the entrenchment of provincial autonomy against the center.
The entrenchment of limits on the provinces, and of the protection of the center’s areas of
competence, can also be tremendously important. Across federal systems probably the most
42
particular institutions may or may not do. It is more like the prohibition on
multiyear defense spending than it is like the existence of state-based
representation in the Senate—there is the potential for interpretive uncertainty
about what qualifies as an expenditure on the armed forces or as interstate
commerce. Guarantees of autonomy remain, however, relatively mechanical.
They do not involve direct dispute over the morally best law or policy. They
redirect interpretive dispute from what should be decided to who should decide; and
they overtly reject any philosopher-king model. (Autonomous provinces will
adopt different laws and policies from one another, so rules will differ depending
on which side of an arbitrary invisible line one stands on—hardly something a
philosopher-king would prescribe.) And, to the degree that they aim to prevent
overcentralization like the rule against perpetual defense spending aims to
prevent overmilitarization, they clearly do so by approximation.
While the
judiciary has some role to perform, it is quite unlike the role of the friend with
the car keys.
important such limits have to do with taxation, spending, and debt. Without credible
prohibitions on provinces pursuing inflationary policies or on their offloading debt onto the
center, responsible budgeting and macroeconomic stability can be badly compromised. The U.S.
Constitution’s prohibition on states printing paper money is the oldest example; a variety of
contemporary constitutional rules enforcing balanced budgeting on provinces or prohibiting
central assumption of provincial debt try to bolster the otherwise-weak credibility of central
governments promising provinces and credit markets that the provinces must stand on their own
fiscal feet. See discussion in Jonathan Rodden, “The Dilemma of Fiscal Federalism: Grants and Fiscal
Performance Around the World,” American Journal of Political Science, Vol. 46, No. 3 (Jul., 2002), pp.
670-687. In the next iteration of this paper, I will discuss entrenched limits on provinces; I am aware that I
need to do so.
43
Contingency and entrenchmemt
I have said much that involves “may” and “might” and contingent and
empirical matters, of course. I do not say that federalist entrenchment is always
constitutionally desirable all things considered—a silly claim anyways, since
federalist entrenchment admits of so much variation that all the possibilities can’t
simultaneously be good ideas. I mean only to suggest that these are the kinds of
considerations that encourage people to adopt decision rules that are not neutral
among outcomes, even in the face of disagreement. If they are sociologically
plausible in a given society, then some degree of federalist entrenchment may
well be desirable.
Citizens of democratic societies will disagree with one another at almost
every level of generality; it has been one of the strengths of Waldron’s work to
rive this point home. Our disagreement about conceptions of the good cannot be
end-run around by imagining agreement about principles of right; we disagree
about those principles, and about their interpretations and meanings even when
we think we agree on the principles. We will disagree about what good policies
are, about what level of government should decide policies, and about how
flexible we should be on both those questions. Constitutional entrenchment
44
cannot be legitimated by imagining universal agreement on procedures or
decision rules, any more than by imagining such consensus on rights.
But it does not follow that we may not enact decision rules, that we may
not make predictions about our own or our successors’ errors and try to stave
them off. Decisions in politics are often not independent of previous decisions; it
would be foolish to insist that in institutional design we pretend that they are.
Forestalling especially serious or likely or irreversible errors, including the errors
of foreclosing future discoveries, is the kind of thing that people do in
institutional design even in the face of disagreement.
Forestalling military
dictatorship, or the shutdown of the policy discovery process provided by
jurisdictional competition, or the elimination of dissent by temporary majorities
or executives—these are moves that can be defended even in the knowledge that
they guarantee some kinds of errors, and even in the knowledge that the agents
they empower—such as judges, but also such as provinces—are not free from
error themselves.
I close with a tentative thought about the implications of this argument for
Waldron’s critique of constitutional entrenchment more generally.
The
distinction I have drawn between relatively mechanical entrenchment and
substantive entrenchment seems to leave Waldron’s critique of the latter intact. I
doubt that matters are quite so simple.
What I have defended is, in part,
45
entrenchment as approximation. I think that the model of the Herculean judge
aiming at morally right answers in each case has had a deforming effect on
Waldron’s theory. He diagnoses the insult to democratic electorates in being told
that, on basic questions of constitutional values, a majority of five out of nine has
profound moral and interpretive insight that trumps the majority of 150 million
plus one out of 300 million.
But perhaps even substantive constitutional
provisions can be understood as compatible with the foreknowledge of error. An
judicially enforceably-entrenched rights provision might have as wide a range of
errors around it as constitutionally unguided legislation—but the midpoint of
those errors could be located in a different place. A democratic majority in a
constitutional moment need not think that it will be mad or drunk in the future
to think that fixing an approximate spot around which errors will fluctuate will
better serve its strongest political values or protect its compromises and
settlements. Fixing that spot does not always give the judiciary final supervision
over it, and the turn to consideration of other actors than courts as constitutional
interpreters has been a salutary one.33
But if the choice is between judicial
supervision and an unmoored provision that risks becoming merely hortatory
because it depends for its interpretation on precisely the political actors it would
33
See, for example, Keith Whittington, Constitutional Construction and Political Foundations of Judicial
Supremacy.
46
have to constrain, then judicial authority need not be a self-insulting choice for
the constitutional enactors to make.